Idea: Reassessing Generative Credit Costs for Firefly Video to Drive Adoption

[Gauher_Jaan3978]

This post is focused on a crucial area for future improvement: the Generative Credit cost model for Firefly Video. My primary goal is to provide candid, constructive feedback that I believe will dramatically increase user adoption and satisfaction with this new feature.

First, I want to express my genuine appreciation. The work you are doing is absolutely groundbreaking. The seamless integration of features like Generative Fill into Photoshop and the artistic quality of your image models has fundamentally changed my creative workflow for the better. As an enthusiastic and appreciative Creative Cloud subscriber, I genuinely root for the success of every new innovation you launch.

The Value Proposition Challenge: I currently have a subscription tier that provides 4,000 Generative Credits per month. Upon the conclusion of the free promotional period for the video model, I noted the following cost structure: The cost for generating a 5-second, 1080p video clip using model "Adobe Firefly Image 5" is 500 Generative Credits.

The Real-World Cost of "Trial and Error" While I understand that generative video is computationally demanding, the current pricing model assumes a high success rate that does not yet exist:

• The Reality of Retries: From my personal experience, achieving a desired, usable result is rarely a "one-shot" process. In my own testing, I found I often needed 18 to 22 retries to get a single satisfactory clip.
• The Multiplier Effect: Because every attempt costs 500 credits—regardless of quality—the effective cost for one usable 5-second video can skyrocket to 10,000 credits or more. This makes the feature financially prohibitive for experimentation.

Comparison to Market Alternatives: in my testing of other commercially available models (such as Google’s Veo 3.1), I observed a significantly higher efficiency rate:

• Efficiency: I was often able to achieve a production-ready 8-second clip (1080p) within just 1 to 3 attempts only.
• Cost: The credit consumption for these successful generations was 400 credits - that is less than the cost of a shorter Firefly clip. The disparity in both cost and success rate makes it difficult to justify using the Firefly model for daily workflows when high-quality alternatives are available at a fraction of the price.

The Commercial Indemnity Factor: I recognize that the higher pricing may be tied to the valuable promise of commercial indemnity (legal protection). However, for a large segment of the community—particularly those of us creating content for social platforms like Instagram—this "premium" protection is often not a priority.

• Social Media Reality: Platforms like Instagram often demonstrate a low concern for copyrighted content when it is clearly made for "fun", "storytelling", and is non-commercial. The primary legal risk arises when a brand attempts to take a piece of UGC and use it in a large-scale public or commercial campaign.
• Limited Scope: Furthermore, even for commercial users, indemnity is not a universal shield against claims based on explicit, infringing user-supplied prompts. For example, if a generation strongly resembles a famous, protected character like Pikachu, the risk of a legal attack from the IP holder remains regardless of Adobe's indemnity.

Paying a substantial premium for a "safety" feature that many of us do not need (or that doesn't fully apply to our specific use caseexample) somewhat feels like a mismatch for the average Creative Cloud user, especially when factoring in the cost of high failure rates.

My Suggestion: I strongly suggest reassessing the credit consumption rate for video generation.

Lowering the cost (or offering a low-cost "preview" mode) would encourage users to experiment more freely. This would not only drive adoption but also provide your team with the massive volume of user feedback needed to refine the model and close the gap with competitors.

I personally really enjoyed this promotion period. We self taught how to work with AI models, and created many videos that I personally find very appealing. For example, take a look at this IG video.

Thank you for listening to your community.
Respectfully,
CQ

[Samarth Syal]

Hi @Gauher_Jaan3978, 

Thank you so much for taking the time to share such thoughtful and detailed feedback, and for being so open about your real-world experience with Firefly Video. It means a lot to the team to hear that Firefly and features like Generative Fill have become such an important part of your creative workflow, and that you’re rooting for these tools to succeed. Your points about the current Generative Credit cost, the impact of retries, and how this compares to other models are completely valid, and it’s clear you’re speaking from deep hands‑on use, not theory. The way you’ve broken down the “trial and error” cost, as well as the nuance around indemnity versus everyday social content, is exactly the kind of perspective that helps shape how we prioritize future improvements. This feedback is being shared with the product team, including your suggestion regarding lower-cost previews and more experimentation-friendly pricing, so it can be considered as we continue to evolve Firefly Video. Thank you again for investing your time, credits, and creativity with us, and for sharing your Instagram example as well. It’s incredibly motivating to see what you’re creating with these tools.

[Scott-name-already-in-use]

This right here nails it:

• The Reality of Retries: From my personal experience, achieving a desired, usable result is rarely a "one-shot" process. In my own testing, I found I often needed 18 to 22 retries to get a single satisfactory clip.
• The Multiplier Effect: Because every attempt costs 500 credits—regardless of quality—the effective cost for one usable 5-second video can skyrocket to 10,000 credits or more. This makes the feature financially prohibitive for experimentation.